- Email: [email protected]
Dynamic power supply by hydrogen bound to a liquid organic hydrogen carrier
Applied Energy 194 (2017) 1–8
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Dynamic power supply by hydrogen bound to a liquid organic hydrogen carrier André Fikrt a,1, Richard Brehmer b,1, Vito-Oronzo Milella c, Karsten Müller a,⇑, Andreas Bösmann b, Patrick Preuster b, Nicolas Alt c, Eberhard Schlücker c, Peter Wasserscheid b, Wolfgang Arlt a a b c
Institute of Separation Science and Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany Institute of Chemical Reaction Engineering, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany Institute for Process Technology and Machinery, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Liquid Organic Hydrogen Carrier can
store hydrogen in a dense form. Dynamics of hydrogen release are
crucial for application in energy storage. Capability of reacting to fast changes in energy demand has been evaluated. The combined system of release and purification unit can react very fast.
a r t i c l e
i n f o
Article history: Received 8 November 2016 Received in revised form 17 February 2017 Accepted 26 February 2017
Keywords: Hydrogen storage LOHC Organic hydrides Dynamic power supply Dehydrogenation Fuel cell
a b s t r a c t Liquid Organic Hydrogen Carriers (LOHCs) are able to store hydrogen in a dense and safe form at ambient conditions. While storage of electrical energy in these carrier systems is one possible and attractive application, the dynamics of the load profile has been believed to represent a major challenge for this storage technology. Conversely, we report here that storage systems based on the LOHC technology are indeed able to deal with significant variations in power demand. This is due to the significant free volume in the LOHC release unit offering the opportunity to handle dynamic behavior by pressure changes. While pressure changes allow quick adaption of the power release on demand, changes in the reactor temperature lead to slow modification of the power output, as demonstrated in this work for hydrogen release from perhydro-dibenzyltoluene (H18-DBT). Ó 2017 Published by Elsevier Ltd.
1. Introduction ⇑ Corresponding author at: Egerlandstr. 3, 91058 Erlangen, Germany. 1
E-mail address: [email protected] (K. Müller). Both authors contributed equally to this work.
http://dx.doi.org/10.1016/j.apenergy.2017.02.070 0306-2619/Ó 2017 Published by Elsevier Ltd.
Liquid Organic Hydrogen Carriers (LOHCs) represent a promising technology for the storage of hydrogen and have attracted
2
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
rising attention in recent years [1–8]. The approach is based on the reversible hydrogenation of an organic carrier compound. Hydrogen is chemically bound to the carrier, which allows safe and dense storage under ambient conditions. A number of different substances such as toluene [9], N-ethylcarbazole [10] or other carbazole derivatives [11] have been described as LOHCs. Due to specific advantageous properties, such as excellent commercial availability, high thermal stability and good hydrogen capacity, the dibenzyltoluene (H0-DBT)/perhydro-dibenzyltoluene (H18DBT) LOHC system is of special relevance for practical and industrial applications (Scheme 1) [12,13]. Although the LOHC technology can be applied for the transport and storage of hydrogen for chemical utilization, the focus of this work is on the storage of energy, e.g. from renewable sources. For high-quality energy storage applications the storage system needs to be able to react sufficiently fast in response to the fluctuating profiles of power production and power demand. While batteries could potentially act as a power buffer, the capability of the LOHC systems for dynamic operation is still an important parameter to determine the required capacity of the battery. Both the hydrogen uptake and hydrogen release units consist of four major elements: a chemical reactor, a hydrogen gas purification system, a heat integration system and an electrochemical conversion system (e.g. electrolysis or fuel cell). Even though the dynamics of fuel cells cannot cope with every load change in the millisecond range [14], they can usually react fast enough on transients for most applications and are significantly faster than typical chemical reactors. The main challenge for increasing the hydrogen output from the dehydrogenation reactor is the fast supply of the required heat into the reactor. Increasing the temperature of the reactor takes at least several minutes, depending on the size of the reactor, the power of the heating system and the heat transport conditions in the reactor. Another option to increase the hydrogen production rate is by increasing the volume flow of hydrogenated LOHC through the reactor. This would increase the hydrogen release at the cost of conversion. In this way, a relatively fast increase of hydrogen output is achievable as pumping performance can be adapted quickly. However, system efficiency and effective storage density would decrease in this way. In previous works several of the current authors have published studies on the physicochemical properties of LOHC materials [11,13,15,16], their phase equilibria [17–19] and their environmental and health impact [5]. A number of theoretical studies by different authors have addressed the efficiency of LOHC based energy storage [20–22] and concepts for the application of the LOHC technology [3,6,23–25]. A large proportion of the work on LOHC published so far focuses on catalysis. Amende et al. [26] studied the mechanism of dehydrogenation on Pd(1 1 1) surfaces. Spectroscopic studies on the dehydrogenation on Pt(1 1 1) have been done by Gleichweit et al. [27]. Matsuda et al. [28] studied the catalysis of dehydrogenation in the liquid phase for a carbazole derivative under ultrahigh vacuum by linking it to an ionic liquid, thus avoiding evaporation. Other works addressed the effects of surface structure on the stability of the LOHC N-ethylcarbazole during dehydrogenation [29]. Studies on catalysis and reaction mechanisms of a variety of potential
LOHC materials have been published in recent years by Dong et al. [30], Mehranfar et al. [31], Brayton and Jensen [32], Papp et al. [33], Do et al. [34], Li et al. [35], Fujita et al. [36] and other authors. Nevertheless, the number of experimental works on LOHC and related hydrogen technologies dealing with actual energy storage systems (i.e. at least hydrogen release plus e.g. fuel cell) is limited. First reports on demonstration plants date back to the 1980s. In these works a release unit for dehydrogenating methyl cyclohexane providing hydrogen to a combustion engine is described [37,38]. More recently two companies have reported the construction of demonstration plants based on toluene/methyl cyclohexane in the Japanese cities of Yokohama [39] and Hitachi [40]. However, the dynamics of electricity supply that can be realized with the LOHC units are not addressed in these publications. To the best of our knowledge there are also no other publications, neither on the dynamics of such systems nor on demonstration plants running with more sophisticated LOHCs. Hence, the dynamics possible with a LOHC system based on dibenzyl toluene should be studied for further improving the understanding of this upcoming technology on a system level. There are some works on metal hydrides touching the issue of dynamics. This is a somewhat related technology for storing hydrogen, since both approaches are based on chemically bound hydrogen and have similar energy demands for hydrogen release. Lototskyy et al. [41] discussed applications for metal hydride based storage and remarked on the respective dynamics. The main limiting factor for fast and dynamic hydrogen release is the heat transfer into the solid material. Laurencelle and Goyette [42] studied the effects of aluminum foam for improving the heat transfer. They concluded that the performance could be significantly improved by introducing these foams. Pasini et al. [43] evaluated metal hydrides concerning the dynamics of hydrogen uptake and concluded that they are insufficient for mobile applications. The solid nature of metal hydrides is one of the major drawbacks for their application for dynamic hydrogen release. Hence, liquid hydrogen carriers might overcome this challenge. This publication elaborates on the possibility for dynamic operation of LOHC-based hydrogen release units through suitable pressure adjustment in the release unit. If the hydrogen production rate does not perfectly match the consumption rate in the fuel cell, elemental hydrogen accumulates in the sections between the reactor and the fuel cell. Thus, the free volume in the connecting pipes of those sections can act as a pressure buffer increasing the ability for dynamics of the system. The flexibility of the option described is limited towards low pressures by the fact that hydrogen needs to flow into the fuel cell. When pressure decreases, the pressure gradient and subsequently the hydrogen flow decrease as well. Therefore, a minimum pressure has to be maintained. A theoretical maximum of the pressure is determined by Le Chatelier’s principle. If the hydrogen pressure is too high dehydrogenation will be limited by its thermodynamical reaction equilibrium. In many cases the maximum pressure will be limited by practical aspects, e.g. by mechanical constraints and safety considerations. However, as will be shown in this paper, only a few hundred millibar of pressure change in the free-volume of the LOHC dehydrogenation unit
Scheme 1. Hydrogen storage and release using the H0-DBT/H18-DBT LOHC system.
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
enable some dynamics of the process and compensates (at least partially) a possible mismatch between the actual dynamics of the reactor and fuel cell. 2. Experimental 2.1. Materials For the dehydrogenation experiments hydrogenated dibenzyltoluol (H18-DBT, hydrogenation grade >95%) from Hydrogenious Technologies GmbH, Germany, was used. All dehydrogenation experiments were carried out using a platinum on alumina support (0.5% Pt loading, Hydrogenious Technologies). The applied heating system was a Unistat TR401 purchased from Peter Huber Kältemaschinenbau GmbH with a maximum power output of 9 kW. 2.2. Experimental setup The hydrogen release unit was designed to represent a LOHC unit for providing energy to a residential building. It was set up as a parallel arrangement of four, horizontal u-tubes placed in a heat exchange box filled with hot heating oil from the heating system. Each tube has a total length of 2.5 m and a diameter of 19 mm. Each reaction tube is completely filled with catalyst. For the herereported experiments only one of the four reaction tubes was used. The dehydrogenation reaction was performed at 1.2–1.5 bar hydrogen pressure and at temperatures between 563 and 593 K. Temperatures and pressures were adjusted depending on the desired hydrogen production rate and degree of dehydrogenation. To minimize heat losses during device operation the heat exchange box was completely covered by mineral wool and all connecting tubing were insulated by glass wool. Downstream to the dehydrogenation reactor, a combined heat integration and cleaning system was applied. The heat integration unit preheats the incoming LOHC with the hot liquid leaving the reactor. Furthermore, a certain amount of the LOHC evaporates in the reactor and in some cases, total vaporization of products and reactants can occur. The product heat consists of sensible heat of hydrogen und LOHC as well as vaporization enthalpy of the LOHC. A countercurrent plate heat exchanger was applied. This device uses the condensation heat for feedstock preheating and serves as partial condenser for initial hydrogen purification. For the final purity, the international standard for hydrogen in automotive fuel cell applications ISO 14687-2 was chosen as a target. It demands a maximum total hydrocarbon amount of 2 ppm [44]. From vapor pressure measurements of dibenzyltoluene, a condensation temperature of 338 K was estimated to be necessary. This corresponds to a partial pressure of 0.3 Pa. To guarantee this temperature at all times, a second partial condenser was set up and operated with cooling water. For continuous separation of gas and liquid phase, a mist separator was designed, based on the Souders–Brown-equation [45]. This unit was combined with a steam trap. It removes liquid LOHC from the gas phase and transfers it to the liquid product tank. The hydrogen gas is purified close to equilibrium conditions. As the Souders–Brown-equation takes only density of phases into account, but neglects e.g. viscosity, the phase separation was tested in advance. The best separation performance was achieved with glass wool mats supported by a perforated stainless steel plate at gas pressures above 110 kPa. Besides the gas-liquid separation for removing/recovering the LOHC from the hydrogen, the phase separation has also to deal with removal of cyclic hydrocarbons. Catalytically cracking of LOHC during the dehydrogenation reaction takes place to a minor degree. This results in the presence of smaller cyclic hydrocarbons
3
with higher vapor pressures. Consequently, there is a risk of exceeding the 2 ppm limit at the given cooling temperature. Preliminary tests indicate that PEM fuel cells do not suffer any obvious damage from contact with LOHC. However, to ensure high purity, as a last safeguard measure, an activated carbon guard bed was installed. Purity of the gas (and composition of the contamination) has been measured between phase separation and the activated carbon guard bed by FTIR spectroscopy (MultiGasTM 2030 FTIR Continuous Gas Analyzer from MKS Instruments). In total, the downstream installations successfully served their purpose in heat recovery and hydrogen purification. In the applied set-up, their volume equaled almost 19 dm3. As a final step, the released und cleaned hydrogen was converted by a 5 kW PEM fuel cell from NextPEM GmbH without additional humidification. The electrical power was fed to the public grid by a DC-AC-converter or was directly consumed by light bulbs. A simplified illustration of the applied LOHC unit, without sensors and control system, is shown in Fig. 1. The fuel cell provides output voltages between 60 and 120 V direct current (DC) depending on electrical load and hydrogen input. The voltage is converted to the German standardized grid voltage of UL = 400 V and frequency of 50 Hz (See ISO IEC 60038, DIN EN 60038:2011). Using a full-bridge converter (KS Schneider Elektronik GmbH) the input voltage is transformed to the output voltage of U2 = 370 V DC. Finally, a line-commutated inverter for three-phase grid connection (Fronius Deutschland GmbH) transfers the DC-voltage into the required alternating current (AC).
3. Dynamics of the release unit The applied dehydrogenation reactor had a total volume of 61 dm3 (steel, catalyst and LOHC) and a total mass of 160 kg. Its heat capacity was about 125 kJ K 1. Consequently, the time required for a cold start is rather long. Heating the reactor from room temperature to the desired reaction temperatures of around 573 K took approximately two hours (Fig. 2). Under the assumption of perfect insulation, the thermostat with a power of about 9 kW, should have brought the reactor temperature to the set point in a little more than one hour. In addition to heat loss to the surroundings, the endothermal effect of hydrogen release further prolongs the time necessary for reaching the set point temperature. Hydrogen release starts at around 523 K (corresponding to about 80 min after start of the heating system). Since hydrogen release is highly endothermal, the reaction consumes a certain share of the heat and thus decelerates the increase in temperature. The LOHC feed flow was started when the reaction temperature reached 573 K. The release of hydrogen from the reactor was a continuous and stationary process after start-up was completed. The purpose of this study was to verify whether the chosen LOHC dehydrogenation set-up would allow dealing with a flexible hydrogen and electricity demand. In general, three parameters can be adjusted to change the power output: (a) the temperature; (b) the LOHC mass flow, and (c) the pressure in the reactor and in the downstream installations. Rising temperature leads to an increase of hydrogen production rate due to faster kinetics (Arrhenius law) and due to a higher driving force of the endothermic dehydrogenation reaction at higher temperature (Le Chatelier’s rule). This effect is depicted in Fig. 3. The target value for temperature was changed by 20 K. The full increase in hydrogen power from 550 Wth,LHV to 820 Wth,LHV takes about 50–60 min. However, hydrogen production rate reacts very fast on increased temperature. The problem is the huge thermal inertia, which causes the temperature to rise rather slowly. Thus, the response time would be far too long to deal with rapid changes
4
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
Fig. 1. Simplified flow chart of the analyzed unit.
Fig. 2. Temperature of the reactor over time during the start-up period.
Fig. 3. Response of hydrogen production rate (expressed in its lower heating value) to an increase of reaction temperature.
in electrical power demand. Because of the rather fast increase of temperature during the first few minutes it takes only 9 min to achieve 66% of that 270 Wth difference and only 14 min to cover 83% of it. Higher conversion is an advantage of increasing the temperature to react on augmented power demand. A higher volume flow of the LOHC causes a decrease of residence time of LOHC in the reactor. This causes a lower degree of dehydrogenation and an increase in the total amount of hydrogen production in the reactor. This is due to a faster reaction rate at low degrees of dehydrogenation. This is shown in Fig. 4. Changing the flow rate (from 30 to 44 ml min–1) results in an adjustment of hydrogen output to a new stable level within
20 min. This is due to the time that is necessary to transfer the system from the previous stationary operation point to the new one. The hydrodynamic residence times for the two flowrates are 4.7 min and 3.3 min respectively. For dynamic operation it is also noteworthy that a significant increase in hydrogen output is already observed within the first 10 min after changing the pumping rate. As shown in Fig. 5, the decrease in conversion with increasing flow rate can be compensated by a simultaneous increase in temperature. If adjusted correctly, the degree of dehydrogenation can be kept constant this way as shown in Fig. 5. Apart from small fluctuations caused by the opposing effects of heating ramp and residence time, the degree of dehydrogenation
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
5
Fig. 4. Conversion and hydrogen release power as a function of time for different pump rates at a reaction temperature of 578 K.
Fig. 5. Simultaneous increase of temperature from 299 to 304 °C and flow rate from 31 to 44 ml min–1. Fig. 6. Exemplary load profile of the fuel cell for a power change by 25%.
can be kept nearly constant by the simultaneous adjustment of temperature and pumping rate. Since an effective use of the LOHC-bound hydrogen is desired in most applications, this operation mode will be most reasonable to realize load changes. While 21 min are needed for the complete transfer into the new stationary operation point, the major part of the change in hydrogen output takes place within the first 10 min after adjusting the parameters. 4. Highly dynamic operation by pressure adjustment The above described approaches to modify the power output are clearly too slow to cope with the dynamics of the fuel cell and the dynamic requirements for a storage unit in a fluctuating electricity grid. From our initial tests of the applied fuel cell with hydrogen 5.0 (supplied by Linde AG) we concluded that the switching time for a jump of 25% in power output of the fuel cell is less than half a second (Fig. 6). Hence, the fuel cell is very suitable for highly dynamic, variable power requirements although it has been reported that such highly dynamic fuel cell operation can result in reduced lifetime of the membrane electrode assembly [46]. Nevertheless, the influence of the fuel cell switching time on the overall dynamics of the LOHC unit can be neglected. The only limitation is an insufficient supply
with hydrogen in times of sudden increase in power demand due to a lack of dynamics in hydrogen release. For further analysis of this aspect, the hydrogen production of the hydrogen release unit was enhanced in two separate experiments: (a) through increasing the volume flow of LOHC by 60%; (b) through increasing the temperature in the reactor by 10 K. These variations result in an additional hydrogen generation by 16–19%, respectively. Thus, it can be concluded that both measures lead to comparable hydrogen production rates. Due to the above mentioned maximum pressure of 150 kPa and the minimum pressure of 110 kPa, the possible buffer effect of a 40 kPa hydrogen pressure range at 25 °C was discussed for the two experiments (Figs. 7 and 8). Figs. 7 and 8 show the dynamics between the starting time zero, where variation of the parameter took place, and the time needed for reaching a new stationary state. With a free gas phase volume of 19 dm3, the pressure change from 150 kPa to 110 kPa is sufficient to increase hydrogen production by 16% (relative to 2.32 kWth) for 10 min. During this time the system can adopt the higher production rate by the increased pumping rate. As can be seen from Fig. 7, the higher productivity based on the higher flow rate sets in after 1 min. 80% of the increase due to higher LOHC flow was achieved within 5.1 min (from starting time zero).
6
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
tion of hydrogen can be decoupled to a certain extent. A simultaneous change of pumping rate and temperature variation and pressure release is desirable to adapt hydrogen release to quickly changing demand.
5. Control of dynamic LOHC systems
Fig. 7. Increased hydrogen generation by increasing the LOHC pumping rate from 5 dm3 h 1 to 8 dm3 h 1.
Fig. 8. Increased hydrogen generation by reactor temperature from 573 K to 583 K.
In contrast, for the change of reaction temperature (Fig. 8), there is a significant delay in the reactor response due the significant thermal inertia of the reactor. This time lag results in the fact that the required production increase of 19% can only be compensated for around three minutes by the pressure change in the system. Only after 10 min the increased hydrogen production sets in. It takes 20 min to reach full hydrogen production after the temperature change. To bridge the thermal inertia in the investigated system, 87 dm3 at 150 kPa or 19 dm3 at 300 kPa maximum pressure of hydrogen would be necessary. To lower thermal inertia and thus reduce the response delay a number of modifications would be possible. Crucial in this context is the ratio between reactor size and maximum heating power. Temperature of small reactor units can be changed faster. However, a lower limit for reactor size is determined by space time yield. If the reactor becomes too small, retention time and thus conversion would drop. The ratio can also be improved by increasing the heating power. Still, heat transport limits the maximum heating power for a given reactor size. Applying heat more directly, e.g. by introducing direct electric heating within the reactor, might further allow for faster increase in reaction temperature. Nevertheless, electrical heating is not a reasonable choice in terms of efficiency. Our study shows that certain fluctuations in power demand can be compensated in LOHC-based hydrogen released systems by the hydrogen pressure in the system. Thus, consumption and produc-
The control of the hydrogen release unit has to ensure that hydrogen release fits to consumer demand over a longer time span. Short time deviations are acceptable, but the upper and lower pressure limit of the apparatuses set constraints. The target value for the pressure is set to 1.3 bar. For controlling hydrogen release and therefore pressure, two parameters (reaction temperature and pump rate) are used. As described above both should be adjusted simultaneously. However, the time span for the hydrogen production to change after the adjustment of temperature and pumping rate differ significantly. In principal, it should be possible to operate the system with pure proportional control. However, this would require a comparatively large range of pumping rates and heating powers for operation. Therefore, it seems more reasonable to apply proportional plus integral control. For setting up such a controller, the delay times of the PPT1 elements (pressure and temperature) and of the PDT1 element (pressure) have to be determined (compare [47]). The delay times have to be distinguished not only between temperature and pumping rate, but also depend on the respective values of these parameters. At a small pumping rate in the range from 10 to 30 ml min 1 the delay time is about 700 s for the system described above. If the pumping rate is increased up to 45 ml min 1 the delay time decreases to about 450 s. Due to the dependence of the delay time on the current value of the pump rate a robust control has to be adaptive. Since the delay times have rather high values and their range is comparatively small, a domain orientated adaptive control with only two domains seems to be sufficient. Identifying the time behavior of hydrogen pressure at the fuel cell input and the power output of the DCDC-converter, the implementation of an abstraction layer is maintainable. In this layer a PID-controller controls the electrical current to the end that the pressure gradient is stabilized to 0. Another issue, being solved, is an increasing power demand at the DCDC-converter exceeding the amount of hydrogen stored in the free volume. A cascading control-level can prevent such a scenario by controlling the LOHC-pump power ratio and finally the hydrogen release of the reactor.
6. Aspects for application of the LOHC technology It was demonstrated that LOHC based energy storage can provide electricity with high flexibility and dynamics. Nevertheless, the characteristics of the dynamic behavior set certain constraints, leading to different suitability for various applications. On the one hand, the systems can respond quickly on changes in the load profile. On the other hand, this fast reaction is provided by a change in pressure, while the chemical process of hydrogen release is rather inert. The LOHC system can cope with changes in power demand up to certain limits. The limit is determined by the free gas volume in the system and the pressure difference from current pressure to the upper pressure limit (if power demand goes down) or to the lower pressure limit respectively (if power demand goes up). If a sudden change in power is very strong, the system can react on it, but is unable to adjust the chemical reaction fast enough. Hence, strong changes in power can also be buffered, but only if they occur as very short peaks.
7
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
Table 1 Share of the LOHC material lost in form of decomposition products slipping through the condensation unit relative to the original LOHC mass and shares of the different decomposition products on the total mass of the light decomposition products (reaction temperature of about 305 °C). Flow rate/ml min 1 Degree of dehydrogenation/– Loss of decomposition products/wt% Share on total of light boiling decomposition products
Methane/wt% Toluene/wt% Benzene/wt% Methyl cyclohexane/wt% Cyclohexane/wt%
17 0.30 0.044 9.5 10.2 34 17 30
30 0.18 0.031 7.8 5.9 19 35 32
44 0.14 0.023 8.5 5.6 17 29 40
The bold number are the main relevant numbers.
Based on the results potential applications can be identified. The LOHC technology seems mainly suited for storage applications with fluctuations of power demand on a minute scale within about 30% of the set point (peaks in the load profile with duration of a few seconds may be higher). Due to the long startup period (and the high energy demand for startup), the technology is mainly suited for applications with permanent energy demand or at least only short periods of shutdown during which temperature can be kept constant. However, in the latter case it should be taken into consideration that some decomposition might occur if the LOHC material is kept inside the reactor for long periods at high temperatures. An interesting application scenario for the LOHC technology would be the storage of energy for residential buildings. The requirements of the respective load profiles could be met with the dynamics reported here for LOHC systems. This holds especially, if the storage is a support for a household with connection to a public power grid. In case of fully self-sufficient, off-grid applications run by renewable energies like PV, a rather large volume of free hydrogen would be required. For such scenarios combination with a battery might be a reasonable choice. The findings of this study, indicate that application of LOHC based hydrogen storage in mobile applications is complicated. In addition to the large mass and volume of the release and purification unit, the dynamics seem insufficient for the operation of cars. For larger systems with less strict dynamic requirements (e.g. trucks) these issues might be less severe. However, a comparatively large free hydrogen volume still would be needed for these applications. 7. Stability and material losses A large share of the free volume that can be used for intermediate storage of free hydrogen is provided by the purification system. Its first step, i.e. the condensation and recycling of LOHC vapor, was able to reduce the loss of LOHC material below detection limit. However, a small loss of material can be observed, due to some decomposition of the LOHC material. The decomposition products are mainly compounds with higher vapor pressures than dibenzyl toluene (e.g. toluene, cyclohexane). Thus, the portion of these compounds, which was not condensed, was higher than for the LOHC material itself. Still, they could be removed almost completely in the activated carbon filter. The material loss due to light boiling decomposition products increases with increasing conversion (see Table 1). For higher conversions decomposition also increases. However, even under total dehydrogenation light decomposition products account for only about 0.2% of the LOHC material. The decomposition products detected in the gas stream are mainly methane, toluene, benzene, methyl cyclohexane and cyclohexane. A greater amount of toluene and benzene was detected when using higher degrees of dehydrogenation (i.e. conversion), while a lower amount of (methyl)
cyclohexane was found. These anticipated findings correspond respectively to the dehydrogenated and to the hydrogenated material. Formation of methane can be attributed mainly to the splitting-off of methyl groups. This splitting-off creates dibenzyl benzene, which has similar properties than dibenzyl toluene and can function as a LOHC material itself. Other products, like benzyl toluene, are also formed but condensed and can function as LOHC materials as well. Therefore, these decomposition products do not necessarily have to be considered as losses. 8. Conclusions Hydrogen stored in the LOHC compound perhydrodibenzyltoluene is released in an endothermic catalytic reaction at temperatures above 523 K. In this contribution, the ability of such a hydrogen release unit for dynamic hydrogen release is evaluated. Due to the significant thermal inertia of the dehydrogenation reactor, the starting-up procedure of a cold dehydrogenation reactor takes a few hours. Once in the period of suitable dehydrogenation temperatures, changes in hydrogen and power release can be induced by temperature and pumping rate variations. The quickest response to transient load profiles is, however, possible by using the free volume of the unit’s purification system as hydrogen buffer. It is demonstrated that a sudden increase in power output from 2.32 kWth to 2.70 kWth (step-change by 16% in power output) can be realized by a pressure change in the 19 dm3 purification unit from 150 kPa to 110 kPa. Such pressure change is able to bridge the delay time for reaching the required, increased hydrogen production in the dehydrogenation reactor by adjustment of the LOHC flow rate. When increasing the flow rate one should consider potential effects on conversion. To avoid a drop in conversion, and thus efficiency, appropriate measures such as increase in temperature or amount of catalyst are required. With the conversions reported in this study an efficiency of about 45% for hydrogen recovery can be reached (losses are to be attributed to combustion of a certain share of the hydrogen for heat production). The operation set points in this study were selected for an analysis of the system dynamics. In a stationary state with high conversion efficiency for hydrogen provision of up to 60% could be achieved. From the results of this work, we conclude that with the right set-up of reactor volume, free buffer volume and fuel cell, dynamic operation of a hot LOHC-based power release unit is indeed possible. Acknowledgments This project has been realized as part of the Bavarian Hydrogen Center and was funded by the Free State of Bavaria. The authors wish to thank Dr. Nollaig Ní Bhriain, Markus Biegel and Alexander Bulgarin for support with experiments and preparation of this manuscript.
8
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
References [1] Biniwale RB, Rayalu S, Devotta S, Ichikawa M. Chemical hydrides: a solution to high capacity hydrogen storage and supply. Int J Hydrogen Energy 2008;33:360–5. [2] Teichmann D, Arlt W, Wasserscheid P, Freymann R. A future energy supply based on Liquid Organic Hydrogen Carriers (LOHC). Energy Environ Sci 2011;4:2767–73. [3] Teichmann D, Stark K, Muller K, Zottl G, Wasserscheid P, Arlt W. Energy storage in residential and commercial buildings via Liquid Organic Hydrogen Carriers (LOHC). Energy Environ Sci 2012;5:9044–54. [4] Müller K, Völkl J, Arlt W. Thermodynamic evaluation of potential organic hydrogen carriers. Energy Technol 2013;1:20–4. [5] Markiewicz M, Zhang YQ, Bosmann A, Bruckner N, Thoming J, Wasserscheid P, et al. Environmental and health impact assessment of Liquid Organic Hydrogen Carrier (LOHC) systems – challenges and preliminary results. Energy Environ Sci 2015;8:1035–45. [6] Geburtig D, Preuster P, Bösmann A, Müller K, Wasserscheid P. Chemical utilization of hydrogen from fluctuating energy sources – catalytic transfer hydrogenation from charged Liquid Organic Hydrogen Carrier systems. Int J Hydrogen Energy 2016;41:1010–7. [7] Eblagon KM, Tam K, Yu KMK, Zhao S-L, Gong X-Q, He H, et al. Study of catalytic sites on ruthenium for hydrogenation of N-ethylcarbazole: implications of hydrogen storage via reversible catalytic hydrogenation. J Phys Chem C 2010;114:9720–30. [8] Sotoodeh F, Huber BJM, Smith KJ. Dehydrogenation kinetics and catalysis of organic heteroaromatics for hydrogen storage. Int J Hydrogen Energy 2012;37:2715–22. [9] Alhumaidan F, Cresswell D, Garforth A. Hydrogen storage in liquid organic hydride: producing hydrogen catalytically from methylcyclohexane. Energy Fuels 2011;25:4217–34. [10] Pez GP, Scott AR, Cooper AC, Cheng H, Wilhelm FC, Abdourazak AH. Hydrogen storage by reversible hydrogenation of pi-conjugated substrates. U.S. patent no. 7,351,395; 2008. [11] Emel’yanenko VN, Varfolomeev MA, Verevkin SP, Stark K, Müller K, Müller M, et al. Hydrogen storage: thermochemical studies of N-alkylcarbazoles and their derivatives as a potential liquid organic hydrogen carriers. J Phys Chem C 2015;119:26381–9. [12] Brückner N, Obesser K, Bösmann A, Teichmann D, Arlt W, Dungs J, et al. Evaluation of industrially applied heat-transfer fluids as liquid organic hydrogen carrier systems. ChemSusChem 2014;7:229–35. [13] Müller K, Stark K, Emel’yanenko VN, Varfolomeev MA, Zaitsau DH, Shoifet E, et al. Liquid organic hydrogen carriers: thermophysical and thermochemical studies of benzyl- and dibenzyl-toluene derivatives. Ind Eng Chem Res 2015;54:7967–76. [14] Caisheng W, Nehrir MH, Shaw SR. Dynamic models and model validation for PEM fuel cells using electrical circuits. IEEE Trans Energy Convers 2005;20:442–51. [15] Stark K, Emel’yanenko VN, Zhabina AA, Varfolomeev MA, Verevkin SP, Müller K, et al. Liquid organic hydrogen carriers: thermophysical and thermochemical studies of carbazole partly and fully hydrogenated derivatives. Ind Eng Chem Res 2015;54:7953–66. [16] Heller A, Rausch MH, Schulz PS, Wasserscheid P, Fröba AP. Binary diffusion coefficients of the liquid organic hydrogen carrier system dibenzyltoluene/ perhydrodibenzyltoluene. J Chem Eng Data 2016;61:504–11. [17] Aslam R, Müller K, Müller M, Koch M, Wasserscheid P, Arlt W. Measurement of hydrogen solubility in potential liquid organic hydrogen carriers. J Chem Eng Data 2016;61:643–9. [18] Aslam R, Müller K, Arlt W. Experimental study of solubility of water in liquid organic hydrogen carriers. J Chem Eng Data 2015;60:1997–2002. [19] Stark K, Keil P, Schug S, Müller K, Wasserscheid P, Arlt W. Melting points of potential liquid organic hydrogen carrier systems consisting of Nalkylcarbazoles. J Chem Eng Data 2016;61:1441–8. [20] Wang H, Zhou X, Ouyang M. Efficiency analysis of novel Liquid Organic Hydrogen Carrier technology and comparison with high pressure storage pathway. Int J Hydrogen Energy 2016;41:18062–71. [21] Müller K, Geng J, Arlt W. Reversible vs. irreversible conversion of hydrogen: how to store energy efficiently? Energy Technol 2013;1:42–7. [22] Adametz P, Pötzinger C, Müller S, Müller K, Preißinger M, Lechner R, et al. Thermodynamic evaluation and carbon footprint analysis of the application of hydrogen-based energy-storage systems in residential buildings. Energy Technol 2016. http://dx.doi.org/10.1002/ente. 201600388.
[23] Krieger C, Müller K, Arlt W. Coupling of a liquid organic hydrogen carrier system with industrial heat. Chem Eng Technol 2016;39:1570–4. [24] Pradhan AU, Shukla A, Pande JV, Karmarkar S, Biniwale RB. A feasibility analysis of hydrogen delivery system using liquid organic hydrides. Int J Hydrogen Energy 2011;36:680–8. [25] Haupt A, Müller K. Integration of a LOHC storage into a heat-controlled CHP system. Energy 2017;117:1123–30. [26] Amende M, Gleichweit C, Xu T, Höfert O, Koch M, Wasserscheid P, et al. Dicyclohexylmethane as a liquid organic hydrogen carrier: a model study on the dehydrogenation mechanism over Pd(1 1 1). Catal Lett 2016;146:851–60. [27] Gleichweit C, Amende M, Höfert O, Xu T, Späth F, Brückner N, et al. Surface reactions of dicyclohexylmethane on Pt(1 1 1). J Phys Chem C 2015;119:20299–311. [28] Matsuda T, Taccardi N, Schwegler J, Wasserscheid P, Steinrück H-P, Maier F. Vacuum surface science meets heterogeneous catalysis: dehydrogenation of a liquid organic hydrogen carrier in the liquid state. ChemPhysChem 2015;16:1873–9. [29] Amende M, Gleichweit C, Schernich S, Höfert O, Lorenz MPA, Zhao W, et al. Size and structure effects controlling the stability of the liquid organic hydrogen carrier dodecahydro-N-ethylcarbazole during dehydrogenation over Pt model catalysts. J Phys Chem Lett 2014;5:1498–504. [30] Dong Y, Yang M, Yang Z, Ke H, Cheng H. Catalytic hydrogenation and dehydrogenation of N-ethylindole as a new heteroaromatic liquid organic hydrogen carrier. Int J Hydrogen Energy 2015;40:10918–22. [31] Mehranfar A, Izadyar M, Esmaeili AA. Hydrogen storage by N-ethylcarbazol as a new liquid organic hydrogen carrier: a DFT study on the mechanism. Int J Hydrogen Energy 2015;40:5797–806. [32] Brayton DF, Jensen CM. Dehydrogenation of pyrrolidine based liquid organic hydrogen carriers by an iridium pincer catalyst, an isothermal kinetic study. Int J Hydrogen Energy 2015;40:16266–70. [33] Papp C, Wasserscheid P, Libuda J, Steinrück H-P. Liquid organic hydrogen carriers: surface science studies of carbazole derivatives. Chem Rec 2014;14:879–96. [34] Do G, Preuster P, Aslam R, Bösmann A, Müller K, Arlt W, et al. Hydrogenation of the liquid organic hydrogen carrier compound dibenzyltoluene – reaction pathway determination by 1H NMR spectroscopy. React Chem Eng 2016;1:313–20. [35] Li X, Tuo Y, Jiang H, Duan X, Yu X, Li P. Engineering Pt/carbonnanofibers/carbon-paper composite towards highly efficient catalyst for hydrogen evolution from liquid organic hydride. Int J Hydrogen Energy 2015;40:12217–26. [36] Fujita K-i, Tanaka Y, Kobayashi M, Yamaguchi R. Homogeneous perdehydrogenation and perhydrogenation of fused bicyclic N-heterocycles catalyzed by iridium complexes bearing a functional bipyridonate ligand. J Am Chem Soc 2014;136:4829–32. [37] Taube M, Rippin D, Knecht W, Hakimifard D, Milisavljevic B, Gruenenfelder N. A prototype truck powered by hydrogen from organic liquid hydrides. Int J Hydrogen Energy 1985;10:595–9. [38] Grünenfelder NF, Schucan TH. Seasonal storage of hydrogen in liquid organic hydrides: description of the second prototype vehicle. Int J Hydrogen Energy 1989;14:579–86. [39] Okada Y, Shimura M. Development of large-scale H2 storage and transportation technology with Liquid Organic Hydrogen Carrier (LOHC). The 21st joint GCC-Japan environment symposium; 2013. p. 5793–803. [40] Emori H, Kawamura T, Ishikawa T, Takamura T. Carbon-hydride energy storage system. Hitachi Rev 2013;62:192. [41] Lototskyy M, Satya Sekhar B, Muthukumar P, Linkov V, Pollet BG. Niche applications of metal hydrides and related thermal management issues. J Alloy Compd 2015;645(Supplement 1). S117–S22. [42] Laurencelle F, Goyette J. Simulation of heat transfer in a metal hydride reactor with aluminium foam. Int J Hydrogen Energy 2007;32:2957–64. [43] Pasini JM, Corgnale C, van Hassel BA, Motyka T, Kumar S, Simmons KL. Metal hydride material requirements for automotive hydrogen storage systems. Int J Hydrogen Energy 2013;38:9755–65. [44] ISO 14687-2:2012. Hydrogen fuel – product specification – Part 2: Proton exchange membrane (PEM) fuel cell applications for road vehicles. [45] Souders M, Brown GG. Design of fractionating columns I. Entrainment and capacity. Ind Eng Chem 1934;26:98–103. [46] Silva RA, Hashimoto T, Thompson GE, Rangel CM. Characterization of MEA degradation for an open air cathode PEM fuel cell. Int J Hydrogen Energy 2012;37:7299–308. [47] Lutz H, Wendt W. Taschenbuch der Regelungstechnik: mit MATLAB und Simulink: Europa Lehrmittel Verlag; 2014.
Contents lists available at ScienceDirect
Applied Energy journal homepage: www.elsevier.com/locate/apenergy
Dynamic power supply by hydrogen bound to a liquid organic hydrogen carrier André Fikrt a,1, Richard Brehmer b,1, Vito-Oronzo Milella c, Karsten Müller a,⇑, Andreas Bösmann b, Patrick Preuster b, Nicolas Alt c, Eberhard Schlücker c, Peter Wasserscheid b, Wolfgang Arlt a a b c
Institute of Separation Science and Technology, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany Institute of Chemical Reaction Engineering, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany Institute for Process Technology and Machinery, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Germany
h i g h l i g h t s
g r a p h i c a l a b s t r a c t
Liquid Organic Hydrogen Carrier can
store hydrogen in a dense form. Dynamics of hydrogen release are
crucial for application in energy storage. Capability of reacting to fast changes in energy demand has been evaluated. The combined system of release and purification unit can react very fast.
a r t i c l e
i n f o
Article history: Received 8 November 2016 Received in revised form 17 February 2017 Accepted 26 February 2017
Keywords: Hydrogen storage LOHC Organic hydrides Dynamic power supply Dehydrogenation Fuel cell
a b s t r a c t Liquid Organic Hydrogen Carriers (LOHCs) are able to store hydrogen in a dense and safe form at ambient conditions. While storage of electrical energy in these carrier systems is one possible and attractive application, the dynamics of the load profile has been believed to represent a major challenge for this storage technology. Conversely, we report here that storage systems based on the LOHC technology are indeed able to deal with significant variations in power demand. This is due to the significant free volume in the LOHC release unit offering the opportunity to handle dynamic behavior by pressure changes. While pressure changes allow quick adaption of the power release on demand, changes in the reactor temperature lead to slow modification of the power output, as demonstrated in this work for hydrogen release from perhydro-dibenzyltoluene (H18-DBT). Ó 2017 Published by Elsevier Ltd.
1. Introduction ⇑ Corresponding author at: Egerlandstr. 3, 91058 Erlangen, Germany. 1
E-mail address: [email protected] (K. Müller). Both authors contributed equally to this work.
http://dx.doi.org/10.1016/j.apenergy.2017.02.070 0306-2619/Ó 2017 Published by Elsevier Ltd.
Liquid Organic Hydrogen Carriers (LOHCs) represent a promising technology for the storage of hydrogen and have attracted
2
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
rising attention in recent years [1–8]. The approach is based on the reversible hydrogenation of an organic carrier compound. Hydrogen is chemically bound to the carrier, which allows safe and dense storage under ambient conditions. A number of different substances such as toluene [9], N-ethylcarbazole [10] or other carbazole derivatives [11] have been described as LOHCs. Due to specific advantageous properties, such as excellent commercial availability, high thermal stability and good hydrogen capacity, the dibenzyltoluene (H0-DBT)/perhydro-dibenzyltoluene (H18DBT) LOHC system is of special relevance for practical and industrial applications (Scheme 1) [12,13]. Although the LOHC technology can be applied for the transport and storage of hydrogen for chemical utilization, the focus of this work is on the storage of energy, e.g. from renewable sources. For high-quality energy storage applications the storage system needs to be able to react sufficiently fast in response to the fluctuating profiles of power production and power demand. While batteries could potentially act as a power buffer, the capability of the LOHC systems for dynamic operation is still an important parameter to determine the required capacity of the battery. Both the hydrogen uptake and hydrogen release units consist of four major elements: a chemical reactor, a hydrogen gas purification system, a heat integration system and an electrochemical conversion system (e.g. electrolysis or fuel cell). Even though the dynamics of fuel cells cannot cope with every load change in the millisecond range [14], they can usually react fast enough on transients for most applications and are significantly faster than typical chemical reactors. The main challenge for increasing the hydrogen output from the dehydrogenation reactor is the fast supply of the required heat into the reactor. Increasing the temperature of the reactor takes at least several minutes, depending on the size of the reactor, the power of the heating system and the heat transport conditions in the reactor. Another option to increase the hydrogen production rate is by increasing the volume flow of hydrogenated LOHC through the reactor. This would increase the hydrogen release at the cost of conversion. In this way, a relatively fast increase of hydrogen output is achievable as pumping performance can be adapted quickly. However, system efficiency and effective storage density would decrease in this way. In previous works several of the current authors have published studies on the physicochemical properties of LOHC materials [11,13,15,16], their phase equilibria [17–19] and their environmental and health impact [5]. A number of theoretical studies by different authors have addressed the efficiency of LOHC based energy storage [20–22] and concepts for the application of the LOHC technology [3,6,23–25]. A large proportion of the work on LOHC published so far focuses on catalysis. Amende et al. [26] studied the mechanism of dehydrogenation on Pd(1 1 1) surfaces. Spectroscopic studies on the dehydrogenation on Pt(1 1 1) have been done by Gleichweit et al. [27]. Matsuda et al. [28] studied the catalysis of dehydrogenation in the liquid phase for a carbazole derivative under ultrahigh vacuum by linking it to an ionic liquid, thus avoiding evaporation. Other works addressed the effects of surface structure on the stability of the LOHC N-ethylcarbazole during dehydrogenation [29]. Studies on catalysis and reaction mechanisms of a variety of potential
LOHC materials have been published in recent years by Dong et al. [30], Mehranfar et al. [31], Brayton and Jensen [32], Papp et al. [33], Do et al. [34], Li et al. [35], Fujita et al. [36] and other authors. Nevertheless, the number of experimental works on LOHC and related hydrogen technologies dealing with actual energy storage systems (i.e. at least hydrogen release plus e.g. fuel cell) is limited. First reports on demonstration plants date back to the 1980s. In these works a release unit for dehydrogenating methyl cyclohexane providing hydrogen to a combustion engine is described [37,38]. More recently two companies have reported the construction of demonstration plants based on toluene/methyl cyclohexane in the Japanese cities of Yokohama [39] and Hitachi [40]. However, the dynamics of electricity supply that can be realized with the LOHC units are not addressed in these publications. To the best of our knowledge there are also no other publications, neither on the dynamics of such systems nor on demonstration plants running with more sophisticated LOHCs. Hence, the dynamics possible with a LOHC system based on dibenzyl toluene should be studied for further improving the understanding of this upcoming technology on a system level. There are some works on metal hydrides touching the issue of dynamics. This is a somewhat related technology for storing hydrogen, since both approaches are based on chemically bound hydrogen and have similar energy demands for hydrogen release. Lototskyy et al. [41] discussed applications for metal hydride based storage and remarked on the respective dynamics. The main limiting factor for fast and dynamic hydrogen release is the heat transfer into the solid material. Laurencelle and Goyette [42] studied the effects of aluminum foam for improving the heat transfer. They concluded that the performance could be significantly improved by introducing these foams. Pasini et al. [43] evaluated metal hydrides concerning the dynamics of hydrogen uptake and concluded that they are insufficient for mobile applications. The solid nature of metal hydrides is one of the major drawbacks for their application for dynamic hydrogen release. Hence, liquid hydrogen carriers might overcome this challenge. This publication elaborates on the possibility for dynamic operation of LOHC-based hydrogen release units through suitable pressure adjustment in the release unit. If the hydrogen production rate does not perfectly match the consumption rate in the fuel cell, elemental hydrogen accumulates in the sections between the reactor and the fuel cell. Thus, the free volume in the connecting pipes of those sections can act as a pressure buffer increasing the ability for dynamics of the system. The flexibility of the option described is limited towards low pressures by the fact that hydrogen needs to flow into the fuel cell. When pressure decreases, the pressure gradient and subsequently the hydrogen flow decrease as well. Therefore, a minimum pressure has to be maintained. A theoretical maximum of the pressure is determined by Le Chatelier’s principle. If the hydrogen pressure is too high dehydrogenation will be limited by its thermodynamical reaction equilibrium. In many cases the maximum pressure will be limited by practical aspects, e.g. by mechanical constraints and safety considerations. However, as will be shown in this paper, only a few hundred millibar of pressure change in the free-volume of the LOHC dehydrogenation unit
Scheme 1. Hydrogen storage and release using the H0-DBT/H18-DBT LOHC system.
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
enable some dynamics of the process and compensates (at least partially) a possible mismatch between the actual dynamics of the reactor and fuel cell. 2. Experimental 2.1. Materials For the dehydrogenation experiments hydrogenated dibenzyltoluol (H18-DBT, hydrogenation grade >95%) from Hydrogenious Technologies GmbH, Germany, was used. All dehydrogenation experiments were carried out using a platinum on alumina support (0.5% Pt loading, Hydrogenious Technologies). The applied heating system was a Unistat TR401 purchased from Peter Huber Kältemaschinenbau GmbH with a maximum power output of 9 kW. 2.2. Experimental setup The hydrogen release unit was designed to represent a LOHC unit for providing energy to a residential building. It was set up as a parallel arrangement of four, horizontal u-tubes placed in a heat exchange box filled with hot heating oil from the heating system. Each tube has a total length of 2.5 m and a diameter of 19 mm. Each reaction tube is completely filled with catalyst. For the herereported experiments only one of the four reaction tubes was used. The dehydrogenation reaction was performed at 1.2–1.5 bar hydrogen pressure and at temperatures between 563 and 593 K. Temperatures and pressures were adjusted depending on the desired hydrogen production rate and degree of dehydrogenation. To minimize heat losses during device operation the heat exchange box was completely covered by mineral wool and all connecting tubing were insulated by glass wool. Downstream to the dehydrogenation reactor, a combined heat integration and cleaning system was applied. The heat integration unit preheats the incoming LOHC with the hot liquid leaving the reactor. Furthermore, a certain amount of the LOHC evaporates in the reactor and in some cases, total vaporization of products and reactants can occur. The product heat consists of sensible heat of hydrogen und LOHC as well as vaporization enthalpy of the LOHC. A countercurrent plate heat exchanger was applied. This device uses the condensation heat for feedstock preheating and serves as partial condenser for initial hydrogen purification. For the final purity, the international standard for hydrogen in automotive fuel cell applications ISO 14687-2 was chosen as a target. It demands a maximum total hydrocarbon amount of 2 ppm [44]. From vapor pressure measurements of dibenzyltoluene, a condensation temperature of 338 K was estimated to be necessary. This corresponds to a partial pressure of 0.3 Pa. To guarantee this temperature at all times, a second partial condenser was set up and operated with cooling water. For continuous separation of gas and liquid phase, a mist separator was designed, based on the Souders–Brown-equation [45]. This unit was combined with a steam trap. It removes liquid LOHC from the gas phase and transfers it to the liquid product tank. The hydrogen gas is purified close to equilibrium conditions. As the Souders–Brown-equation takes only density of phases into account, but neglects e.g. viscosity, the phase separation was tested in advance. The best separation performance was achieved with glass wool mats supported by a perforated stainless steel plate at gas pressures above 110 kPa. Besides the gas-liquid separation for removing/recovering the LOHC from the hydrogen, the phase separation has also to deal with removal of cyclic hydrocarbons. Catalytically cracking of LOHC during the dehydrogenation reaction takes place to a minor degree. This results in the presence of smaller cyclic hydrocarbons
3
with higher vapor pressures. Consequently, there is a risk of exceeding the 2 ppm limit at the given cooling temperature. Preliminary tests indicate that PEM fuel cells do not suffer any obvious damage from contact with LOHC. However, to ensure high purity, as a last safeguard measure, an activated carbon guard bed was installed. Purity of the gas (and composition of the contamination) has been measured between phase separation and the activated carbon guard bed by FTIR spectroscopy (MultiGasTM 2030 FTIR Continuous Gas Analyzer from MKS Instruments). In total, the downstream installations successfully served their purpose in heat recovery and hydrogen purification. In the applied set-up, their volume equaled almost 19 dm3. As a final step, the released und cleaned hydrogen was converted by a 5 kW PEM fuel cell from NextPEM GmbH without additional humidification. The electrical power was fed to the public grid by a DC-AC-converter or was directly consumed by light bulbs. A simplified illustration of the applied LOHC unit, without sensors and control system, is shown in Fig. 1. The fuel cell provides output voltages between 60 and 120 V direct current (DC) depending on electrical load and hydrogen input. The voltage is converted to the German standardized grid voltage of UL = 400 V and frequency of 50 Hz (See ISO IEC 60038, DIN EN 60038:2011). Using a full-bridge converter (KS Schneider Elektronik GmbH) the input voltage is transformed to the output voltage of U2 = 370 V DC. Finally, a line-commutated inverter for three-phase grid connection (Fronius Deutschland GmbH) transfers the DC-voltage into the required alternating current (AC).
3. Dynamics of the release unit The applied dehydrogenation reactor had a total volume of 61 dm3 (steel, catalyst and LOHC) and a total mass of 160 kg. Its heat capacity was about 125 kJ K 1. Consequently, the time required for a cold start is rather long. Heating the reactor from room temperature to the desired reaction temperatures of around 573 K took approximately two hours (Fig. 2). Under the assumption of perfect insulation, the thermostat with a power of about 9 kW, should have brought the reactor temperature to the set point in a little more than one hour. In addition to heat loss to the surroundings, the endothermal effect of hydrogen release further prolongs the time necessary for reaching the set point temperature. Hydrogen release starts at around 523 K (corresponding to about 80 min after start of the heating system). Since hydrogen release is highly endothermal, the reaction consumes a certain share of the heat and thus decelerates the increase in temperature. The LOHC feed flow was started when the reaction temperature reached 573 K. The release of hydrogen from the reactor was a continuous and stationary process after start-up was completed. The purpose of this study was to verify whether the chosen LOHC dehydrogenation set-up would allow dealing with a flexible hydrogen and electricity demand. In general, three parameters can be adjusted to change the power output: (a) the temperature; (b) the LOHC mass flow, and (c) the pressure in the reactor and in the downstream installations. Rising temperature leads to an increase of hydrogen production rate due to faster kinetics (Arrhenius law) and due to a higher driving force of the endothermic dehydrogenation reaction at higher temperature (Le Chatelier’s rule). This effect is depicted in Fig. 3. The target value for temperature was changed by 20 K. The full increase in hydrogen power from 550 Wth,LHV to 820 Wth,LHV takes about 50–60 min. However, hydrogen production rate reacts very fast on increased temperature. The problem is the huge thermal inertia, which causes the temperature to rise rather slowly. Thus, the response time would be far too long to deal with rapid changes
4
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
Fig. 1. Simplified flow chart of the analyzed unit.
Fig. 2. Temperature of the reactor over time during the start-up period.
Fig. 3. Response of hydrogen production rate (expressed in its lower heating value) to an increase of reaction temperature.
in electrical power demand. Because of the rather fast increase of temperature during the first few minutes it takes only 9 min to achieve 66% of that 270 Wth difference and only 14 min to cover 83% of it. Higher conversion is an advantage of increasing the temperature to react on augmented power demand. A higher volume flow of the LOHC causes a decrease of residence time of LOHC in the reactor. This causes a lower degree of dehydrogenation and an increase in the total amount of hydrogen production in the reactor. This is due to a faster reaction rate at low degrees of dehydrogenation. This is shown in Fig. 4. Changing the flow rate (from 30 to 44 ml min–1) results in an adjustment of hydrogen output to a new stable level within
20 min. This is due to the time that is necessary to transfer the system from the previous stationary operation point to the new one. The hydrodynamic residence times for the two flowrates are 4.7 min and 3.3 min respectively. For dynamic operation it is also noteworthy that a significant increase in hydrogen output is already observed within the first 10 min after changing the pumping rate. As shown in Fig. 5, the decrease in conversion with increasing flow rate can be compensated by a simultaneous increase in temperature. If adjusted correctly, the degree of dehydrogenation can be kept constant this way as shown in Fig. 5. Apart from small fluctuations caused by the opposing effects of heating ramp and residence time, the degree of dehydrogenation
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
5
Fig. 4. Conversion and hydrogen release power as a function of time for different pump rates at a reaction temperature of 578 K.
Fig. 5. Simultaneous increase of temperature from 299 to 304 °C and flow rate from 31 to 44 ml min–1. Fig. 6. Exemplary load profile of the fuel cell for a power change by 25%.
can be kept nearly constant by the simultaneous adjustment of temperature and pumping rate. Since an effective use of the LOHC-bound hydrogen is desired in most applications, this operation mode will be most reasonable to realize load changes. While 21 min are needed for the complete transfer into the new stationary operation point, the major part of the change in hydrogen output takes place within the first 10 min after adjusting the parameters. 4. Highly dynamic operation by pressure adjustment The above described approaches to modify the power output are clearly too slow to cope with the dynamics of the fuel cell and the dynamic requirements for a storage unit in a fluctuating electricity grid. From our initial tests of the applied fuel cell with hydrogen 5.0 (supplied by Linde AG) we concluded that the switching time for a jump of 25% in power output of the fuel cell is less than half a second (Fig. 6). Hence, the fuel cell is very suitable for highly dynamic, variable power requirements although it has been reported that such highly dynamic fuel cell operation can result in reduced lifetime of the membrane electrode assembly [46]. Nevertheless, the influence of the fuel cell switching time on the overall dynamics of the LOHC unit can be neglected. The only limitation is an insufficient supply
with hydrogen in times of sudden increase in power demand due to a lack of dynamics in hydrogen release. For further analysis of this aspect, the hydrogen production of the hydrogen release unit was enhanced in two separate experiments: (a) through increasing the volume flow of LOHC by 60%; (b) through increasing the temperature in the reactor by 10 K. These variations result in an additional hydrogen generation by 16–19%, respectively. Thus, it can be concluded that both measures lead to comparable hydrogen production rates. Due to the above mentioned maximum pressure of 150 kPa and the minimum pressure of 110 kPa, the possible buffer effect of a 40 kPa hydrogen pressure range at 25 °C was discussed for the two experiments (Figs. 7 and 8). Figs. 7 and 8 show the dynamics between the starting time zero, where variation of the parameter took place, and the time needed for reaching a new stationary state. With a free gas phase volume of 19 dm3, the pressure change from 150 kPa to 110 kPa is sufficient to increase hydrogen production by 16% (relative to 2.32 kWth) for 10 min. During this time the system can adopt the higher production rate by the increased pumping rate. As can be seen from Fig. 7, the higher productivity based on the higher flow rate sets in after 1 min. 80% of the increase due to higher LOHC flow was achieved within 5.1 min (from starting time zero).
6
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
tion of hydrogen can be decoupled to a certain extent. A simultaneous change of pumping rate and temperature variation and pressure release is desirable to adapt hydrogen release to quickly changing demand.
5. Control of dynamic LOHC systems
Fig. 7. Increased hydrogen generation by increasing the LOHC pumping rate from 5 dm3 h 1 to 8 dm3 h 1.
Fig. 8. Increased hydrogen generation by reactor temperature from 573 K to 583 K.
In contrast, for the change of reaction temperature (Fig. 8), there is a significant delay in the reactor response due the significant thermal inertia of the reactor. This time lag results in the fact that the required production increase of 19% can only be compensated for around three minutes by the pressure change in the system. Only after 10 min the increased hydrogen production sets in. It takes 20 min to reach full hydrogen production after the temperature change. To bridge the thermal inertia in the investigated system, 87 dm3 at 150 kPa or 19 dm3 at 300 kPa maximum pressure of hydrogen would be necessary. To lower thermal inertia and thus reduce the response delay a number of modifications would be possible. Crucial in this context is the ratio between reactor size and maximum heating power. Temperature of small reactor units can be changed faster. However, a lower limit for reactor size is determined by space time yield. If the reactor becomes too small, retention time and thus conversion would drop. The ratio can also be improved by increasing the heating power. Still, heat transport limits the maximum heating power for a given reactor size. Applying heat more directly, e.g. by introducing direct electric heating within the reactor, might further allow for faster increase in reaction temperature. Nevertheless, electrical heating is not a reasonable choice in terms of efficiency. Our study shows that certain fluctuations in power demand can be compensated in LOHC-based hydrogen released systems by the hydrogen pressure in the system. Thus, consumption and produc-
The control of the hydrogen release unit has to ensure that hydrogen release fits to consumer demand over a longer time span. Short time deviations are acceptable, but the upper and lower pressure limit of the apparatuses set constraints. The target value for the pressure is set to 1.3 bar. For controlling hydrogen release and therefore pressure, two parameters (reaction temperature and pump rate) are used. As described above both should be adjusted simultaneously. However, the time span for the hydrogen production to change after the adjustment of temperature and pumping rate differ significantly. In principal, it should be possible to operate the system with pure proportional control. However, this would require a comparatively large range of pumping rates and heating powers for operation. Therefore, it seems more reasonable to apply proportional plus integral control. For setting up such a controller, the delay times of the PPT1 elements (pressure and temperature) and of the PDT1 element (pressure) have to be determined (compare [47]). The delay times have to be distinguished not only between temperature and pumping rate, but also depend on the respective values of these parameters. At a small pumping rate in the range from 10 to 30 ml min 1 the delay time is about 700 s for the system described above. If the pumping rate is increased up to 45 ml min 1 the delay time decreases to about 450 s. Due to the dependence of the delay time on the current value of the pump rate a robust control has to be adaptive. Since the delay times have rather high values and their range is comparatively small, a domain orientated adaptive control with only two domains seems to be sufficient. Identifying the time behavior of hydrogen pressure at the fuel cell input and the power output of the DCDC-converter, the implementation of an abstraction layer is maintainable. In this layer a PID-controller controls the electrical current to the end that the pressure gradient is stabilized to 0. Another issue, being solved, is an increasing power demand at the DCDC-converter exceeding the amount of hydrogen stored in the free volume. A cascading control-level can prevent such a scenario by controlling the LOHC-pump power ratio and finally the hydrogen release of the reactor.
6. Aspects for application of the LOHC technology It was demonstrated that LOHC based energy storage can provide electricity with high flexibility and dynamics. Nevertheless, the characteristics of the dynamic behavior set certain constraints, leading to different suitability for various applications. On the one hand, the systems can respond quickly on changes in the load profile. On the other hand, this fast reaction is provided by a change in pressure, while the chemical process of hydrogen release is rather inert. The LOHC system can cope with changes in power demand up to certain limits. The limit is determined by the free gas volume in the system and the pressure difference from current pressure to the upper pressure limit (if power demand goes down) or to the lower pressure limit respectively (if power demand goes up). If a sudden change in power is very strong, the system can react on it, but is unable to adjust the chemical reaction fast enough. Hence, strong changes in power can also be buffered, but only if they occur as very short peaks.
7
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
Table 1 Share of the LOHC material lost in form of decomposition products slipping through the condensation unit relative to the original LOHC mass and shares of the different decomposition products on the total mass of the light decomposition products (reaction temperature of about 305 °C). Flow rate/ml min 1 Degree of dehydrogenation/– Loss of decomposition products/wt% Share on total of light boiling decomposition products
Methane/wt% Toluene/wt% Benzene/wt% Methyl cyclohexane/wt% Cyclohexane/wt%
17 0.30 0.044 9.5 10.2 34 17 30
30 0.18 0.031 7.8 5.9 19 35 32
44 0.14 0.023 8.5 5.6 17 29 40
The bold number are the main relevant numbers.
Based on the results potential applications can be identified. The LOHC technology seems mainly suited for storage applications with fluctuations of power demand on a minute scale within about 30% of the set point (peaks in the load profile with duration of a few seconds may be higher). Due to the long startup period (and the high energy demand for startup), the technology is mainly suited for applications with permanent energy demand or at least only short periods of shutdown during which temperature can be kept constant. However, in the latter case it should be taken into consideration that some decomposition might occur if the LOHC material is kept inside the reactor for long periods at high temperatures. An interesting application scenario for the LOHC technology would be the storage of energy for residential buildings. The requirements of the respective load profiles could be met with the dynamics reported here for LOHC systems. This holds especially, if the storage is a support for a household with connection to a public power grid. In case of fully self-sufficient, off-grid applications run by renewable energies like PV, a rather large volume of free hydrogen would be required. For such scenarios combination with a battery might be a reasonable choice. The findings of this study, indicate that application of LOHC based hydrogen storage in mobile applications is complicated. In addition to the large mass and volume of the release and purification unit, the dynamics seem insufficient for the operation of cars. For larger systems with less strict dynamic requirements (e.g. trucks) these issues might be less severe. However, a comparatively large free hydrogen volume still would be needed for these applications. 7. Stability and material losses A large share of the free volume that can be used for intermediate storage of free hydrogen is provided by the purification system. Its first step, i.e. the condensation and recycling of LOHC vapor, was able to reduce the loss of LOHC material below detection limit. However, a small loss of material can be observed, due to some decomposition of the LOHC material. The decomposition products are mainly compounds with higher vapor pressures than dibenzyl toluene (e.g. toluene, cyclohexane). Thus, the portion of these compounds, which was not condensed, was higher than for the LOHC material itself. Still, they could be removed almost completely in the activated carbon filter. The material loss due to light boiling decomposition products increases with increasing conversion (see Table 1). For higher conversions decomposition also increases. However, even under total dehydrogenation light decomposition products account for only about 0.2% of the LOHC material. The decomposition products detected in the gas stream are mainly methane, toluene, benzene, methyl cyclohexane and cyclohexane. A greater amount of toluene and benzene was detected when using higher degrees of dehydrogenation (i.e. conversion), while a lower amount of (methyl)
cyclohexane was found. These anticipated findings correspond respectively to the dehydrogenated and to the hydrogenated material. Formation of methane can be attributed mainly to the splitting-off of methyl groups. This splitting-off creates dibenzyl benzene, which has similar properties than dibenzyl toluene and can function as a LOHC material itself. Other products, like benzyl toluene, are also formed but condensed and can function as LOHC materials as well. Therefore, these decomposition products do not necessarily have to be considered as losses. 8. Conclusions Hydrogen stored in the LOHC compound perhydrodibenzyltoluene is released in an endothermic catalytic reaction at temperatures above 523 K. In this contribution, the ability of such a hydrogen release unit for dynamic hydrogen release is evaluated. Due to the significant thermal inertia of the dehydrogenation reactor, the starting-up procedure of a cold dehydrogenation reactor takes a few hours. Once in the period of suitable dehydrogenation temperatures, changes in hydrogen and power release can be induced by temperature and pumping rate variations. The quickest response to transient load profiles is, however, possible by using the free volume of the unit’s purification system as hydrogen buffer. It is demonstrated that a sudden increase in power output from 2.32 kWth to 2.70 kWth (step-change by 16% in power output) can be realized by a pressure change in the 19 dm3 purification unit from 150 kPa to 110 kPa. Such pressure change is able to bridge the delay time for reaching the required, increased hydrogen production in the dehydrogenation reactor by adjustment of the LOHC flow rate. When increasing the flow rate one should consider potential effects on conversion. To avoid a drop in conversion, and thus efficiency, appropriate measures such as increase in temperature or amount of catalyst are required. With the conversions reported in this study an efficiency of about 45% for hydrogen recovery can be reached (losses are to be attributed to combustion of a certain share of the hydrogen for heat production). The operation set points in this study were selected for an analysis of the system dynamics. In a stationary state with high conversion efficiency for hydrogen provision of up to 60% could be achieved. From the results of this work, we conclude that with the right set-up of reactor volume, free buffer volume and fuel cell, dynamic operation of a hot LOHC-based power release unit is indeed possible. Acknowledgments This project has been realized as part of the Bavarian Hydrogen Center and was funded by the Free State of Bavaria. The authors wish to thank Dr. Nollaig Ní Bhriain, Markus Biegel and Alexander Bulgarin for support with experiments and preparation of this manuscript.
8
A. Fikrt et al. / Applied Energy 194 (2017) 1–8
References [1] Biniwale RB, Rayalu S, Devotta S, Ichikawa M. Chemical hydrides: a solution to high capacity hydrogen storage and supply. Int J Hydrogen Energy 2008;33:360–5. [2] Teichmann D, Arlt W, Wasserscheid P, Freymann R. A future energy supply based on Liquid Organic Hydrogen Carriers (LOHC). Energy Environ Sci 2011;4:2767–73. [3] Teichmann D, Stark K, Muller K, Zottl G, Wasserscheid P, Arlt W. Energy storage in residential and commercial buildings via Liquid Organic Hydrogen Carriers (LOHC). Energy Environ Sci 2012;5:9044–54. [4] Müller K, Völkl J, Arlt W. Thermodynamic evaluation of potential organic hydrogen carriers. Energy Technol 2013;1:20–4. [5] Markiewicz M, Zhang YQ, Bosmann A, Bruckner N, Thoming J, Wasserscheid P, et al. Environmental and health impact assessment of Liquid Organic Hydrogen Carrier (LOHC) systems – challenges and preliminary results. Energy Environ Sci 2015;8:1035–45. [6] Geburtig D, Preuster P, Bösmann A, Müller K, Wasserscheid P. Chemical utilization of hydrogen from fluctuating energy sources – catalytic transfer hydrogenation from charged Liquid Organic Hydrogen Carrier systems. Int J Hydrogen Energy 2016;41:1010–7. [7] Eblagon KM, Tam K, Yu KMK, Zhao S-L, Gong X-Q, He H, et al. Study of catalytic sites on ruthenium for hydrogenation of N-ethylcarbazole: implications of hydrogen storage via reversible catalytic hydrogenation. J Phys Chem C 2010;114:9720–30. [8] Sotoodeh F, Huber BJM, Smith KJ. Dehydrogenation kinetics and catalysis of organic heteroaromatics for hydrogen storage. Int J Hydrogen Energy 2012;37:2715–22. [9] Alhumaidan F, Cresswell D, Garforth A. Hydrogen storage in liquid organic hydride: producing hydrogen catalytically from methylcyclohexane. Energy Fuels 2011;25:4217–34. [10] Pez GP, Scott AR, Cooper AC, Cheng H, Wilhelm FC, Abdourazak AH. Hydrogen storage by reversible hydrogenation of pi-conjugated substrates. U.S. patent no. 7,351,395; 2008. [11] Emel’yanenko VN, Varfolomeev MA, Verevkin SP, Stark K, Müller K, Müller M, et al. Hydrogen storage: thermochemical studies of N-alkylcarbazoles and their derivatives as a potential liquid organic hydrogen carriers. J Phys Chem C 2015;119:26381–9. [12] Brückner N, Obesser K, Bösmann A, Teichmann D, Arlt W, Dungs J, et al. Evaluation of industrially applied heat-transfer fluids as liquid organic hydrogen carrier systems. ChemSusChem 2014;7:229–35. [13] Müller K, Stark K, Emel’yanenko VN, Varfolomeev MA, Zaitsau DH, Shoifet E, et al. Liquid organic hydrogen carriers: thermophysical and thermochemical studies of benzyl- and dibenzyl-toluene derivatives. Ind Eng Chem Res 2015;54:7967–76. [14] Caisheng W, Nehrir MH, Shaw SR. Dynamic models and model validation for PEM fuel cells using electrical circuits. IEEE Trans Energy Convers 2005;20:442–51. [15] Stark K, Emel’yanenko VN, Zhabina AA, Varfolomeev MA, Verevkin SP, Müller K, et al. Liquid organic hydrogen carriers: thermophysical and thermochemical studies of carbazole partly and fully hydrogenated derivatives. Ind Eng Chem Res 2015;54:7953–66. [16] Heller A, Rausch MH, Schulz PS, Wasserscheid P, Fröba AP. Binary diffusion coefficients of the liquid organic hydrogen carrier system dibenzyltoluene/ perhydrodibenzyltoluene. J Chem Eng Data 2016;61:504–11. [17] Aslam R, Müller K, Müller M, Koch M, Wasserscheid P, Arlt W. Measurement of hydrogen solubility in potential liquid organic hydrogen carriers. J Chem Eng Data 2016;61:643–9. [18] Aslam R, Müller K, Arlt W. Experimental study of solubility of water in liquid organic hydrogen carriers. J Chem Eng Data 2015;60:1997–2002. [19] Stark K, Keil P, Schug S, Müller K, Wasserscheid P, Arlt W. Melting points of potential liquid organic hydrogen carrier systems consisting of Nalkylcarbazoles. J Chem Eng Data 2016;61:1441–8. [20] Wang H, Zhou X, Ouyang M. Efficiency analysis of novel Liquid Organic Hydrogen Carrier technology and comparison with high pressure storage pathway. Int J Hydrogen Energy 2016;41:18062–71. [21] Müller K, Geng J, Arlt W. Reversible vs. irreversible conversion of hydrogen: how to store energy efficiently? Energy Technol 2013;1:42–7. [22] Adametz P, Pötzinger C, Müller S, Müller K, Preißinger M, Lechner R, et al. Thermodynamic evaluation and carbon footprint analysis of the application of hydrogen-based energy-storage systems in residential buildings. Energy Technol 2016. http://dx.doi.org/10.1002/ente. 201600388.
[23] Krieger C, Müller K, Arlt W. Coupling of a liquid organic hydrogen carrier system with industrial heat. Chem Eng Technol 2016;39:1570–4. [24] Pradhan AU, Shukla A, Pande JV, Karmarkar S, Biniwale RB. A feasibility analysis of hydrogen delivery system using liquid organic hydrides. Int J Hydrogen Energy 2011;36:680–8. [25] Haupt A, Müller K. Integration of a LOHC storage into a heat-controlled CHP system. Energy 2017;117:1123–30. [26] Amende M, Gleichweit C, Xu T, Höfert O, Koch M, Wasserscheid P, et al. Dicyclohexylmethane as a liquid organic hydrogen carrier: a model study on the dehydrogenation mechanism over Pd(1 1 1). Catal Lett 2016;146:851–60. [27] Gleichweit C, Amende M, Höfert O, Xu T, Späth F, Brückner N, et al. Surface reactions of dicyclohexylmethane on Pt(1 1 1). J Phys Chem C 2015;119:20299–311. [28] Matsuda T, Taccardi N, Schwegler J, Wasserscheid P, Steinrück H-P, Maier F. Vacuum surface science meets heterogeneous catalysis: dehydrogenation of a liquid organic hydrogen carrier in the liquid state. ChemPhysChem 2015;16:1873–9. [29] Amende M, Gleichweit C, Schernich S, Höfert O, Lorenz MPA, Zhao W, et al. Size and structure effects controlling the stability of the liquid organic hydrogen carrier dodecahydro-N-ethylcarbazole during dehydrogenation over Pt model catalysts. J Phys Chem Lett 2014;5:1498–504. [30] Dong Y, Yang M, Yang Z, Ke H, Cheng H. Catalytic hydrogenation and dehydrogenation of N-ethylindole as a new heteroaromatic liquid organic hydrogen carrier. Int J Hydrogen Energy 2015;40:10918–22. [31] Mehranfar A, Izadyar M, Esmaeili AA. Hydrogen storage by N-ethylcarbazol as a new liquid organic hydrogen carrier: a DFT study on the mechanism. Int J Hydrogen Energy 2015;40:5797–806. [32] Brayton DF, Jensen CM. Dehydrogenation of pyrrolidine based liquid organic hydrogen carriers by an iridium pincer catalyst, an isothermal kinetic study. Int J Hydrogen Energy 2015;40:16266–70. [33] Papp C, Wasserscheid P, Libuda J, Steinrück H-P. Liquid organic hydrogen carriers: surface science studies of carbazole derivatives. Chem Rec 2014;14:879–96. [34] Do G, Preuster P, Aslam R, Bösmann A, Müller K, Arlt W, et al. Hydrogenation of the liquid organic hydrogen carrier compound dibenzyltoluene – reaction pathway determination by 1H NMR spectroscopy. React Chem Eng 2016;1:313–20. [35] Li X, Tuo Y, Jiang H, Duan X, Yu X, Li P. Engineering Pt/carbonnanofibers/carbon-paper composite towards highly efficient catalyst for hydrogen evolution from liquid organic hydride. Int J Hydrogen Energy 2015;40:12217–26. [36] Fujita K-i, Tanaka Y, Kobayashi M, Yamaguchi R. Homogeneous perdehydrogenation and perhydrogenation of fused bicyclic N-heterocycles catalyzed by iridium complexes bearing a functional bipyridonate ligand. J Am Chem Soc 2014;136:4829–32. [37] Taube M, Rippin D, Knecht W, Hakimifard D, Milisavljevic B, Gruenenfelder N. A prototype truck powered by hydrogen from organic liquid hydrides. Int J Hydrogen Energy 1985;10:595–9. [38] Grünenfelder NF, Schucan TH. Seasonal storage of hydrogen in liquid organic hydrides: description of the second prototype vehicle. Int J Hydrogen Energy 1989;14:579–86. [39] Okada Y, Shimura M. Development of large-scale H2 storage and transportation technology with Liquid Organic Hydrogen Carrier (LOHC). The 21st joint GCC-Japan environment symposium; 2013. p. 5793–803. [40] Emori H, Kawamura T, Ishikawa T, Takamura T. Carbon-hydride energy storage system. Hitachi Rev 2013;62:192. [41] Lototskyy M, Satya Sekhar B, Muthukumar P, Linkov V, Pollet BG. Niche applications of metal hydrides and related thermal management issues. J Alloy Compd 2015;645(Supplement 1). S117–S22. [42] Laurencelle F, Goyette J. Simulation of heat transfer in a metal hydride reactor with aluminium foam. Int J Hydrogen Energy 2007;32:2957–64. [43] Pasini JM, Corgnale C, van Hassel BA, Motyka T, Kumar S, Simmons KL. Metal hydride material requirements for automotive hydrogen storage systems. Int J Hydrogen Energy 2013;38:9755–65. [44] ISO 14687-2:2012. Hydrogen fuel – product specification – Part 2: Proton exchange membrane (PEM) fuel cell applications for road vehicles. [45] Souders M, Brown GG. Design of fractionating columns I. Entrainment and capacity. Ind Eng Chem 1934;26:98–103. [46] Silva RA, Hashimoto T, Thompson GE, Rangel CM. Characterization of MEA degradation for an open air cathode PEM fuel cell. Int J Hydrogen Energy 2012;37:7299–308. [47] Lutz H, Wendt W. Taschenbuch der Regelungstechnik: mit MATLAB und Simulink: Europa Lehrmittel Verlag; 2014.